main (PDF)

Network Selection Algorithms for
Multi-Homed mobile Terminals in
a Heterogeneous Network Using
Utility-based MADM and Mobile
Terminal Movement Prediction

by Jiamo Liu
Prepared for O. Falowo
Department of Electrical Engineering
University of Cape Town
Submitted to the Department of Electrical Engineering at the University of Cape
Town in partial fulfilment of the academic requirements for a Bachelor of Science
degree in Electrical and Computer Engineering.
November 2016

Declaration
1. I know that plagiarism is wrong. Plagiarism is to use another’s work and
pretend that it is one’s own.
2. I have used the IEEE convention for citation and referencing. Each
contribution to, and quotation in, this report from the work(s) of other
people has been attributed, and has been cited and referenced.
3. This report is my own work.
4. I have not allowed, and will not allow, anyone to copy my work with the
intention of passing it off as their own work or part thereof.

Signature:

Date:

...........................
Jiamo Liu
13/11/2016

Abstract
As the number of network subscribers increases, it has become a very
challenging task to satisfy the ever-increasing demand of bandwidth by mobile
terminals using current infrastructures. Multi-homed terminals are proposed as
one of the possible solutions because of its bandwidth-aggregation feature, thus this
paper proposes a network selection algorithm for multi-homed mobile terminals
using movement prediction and MADM.
In this paper, network and movement models are establish in order to facilitate
the simulation. The algorithm predicts the positions of the user with Kalman
Filter, and then establish a list of RATs that are qualified as candidates using
information of predictions, finally the algorithm iterates through all combinations
of the candidates and evaluate the suitability of each combination based on three
attributes, namely monetary cost, power dissipation and bandwidth utility. Once
the most suitable set of interfaces are found, the user will switch his connections
to the most suitable combination of interfaces if the dead time has passed, in order
to alleviate the ping-pong effect.
In addition to making logical network interface selection decisions, the
simulation of the algorithm has also suggested that the prediction mechanism is
able to reduce the number of hand-overs caused by ping-pong effect and ensure the
ubiquitous connection within the prediction window period if the predictions are
reasonably accurate. Furthermore it is also observed that prediction mechanism
enhances the rationality of the decision making process in certain scenarios.

Contents
List of Figures

v

List of Tables

vi

1 Introduction
1.1 Background to the study . . . . . .
1.2 Objectives of this study . . . . . .
1.2.1 Problems to be investigated
1.2.2 Purpose of the study . . . .
1.3 Scope and limitations . . . . . . . .
1.4 Plan of development . . . . . . . .

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

2 Literature Review
2.1 Development of wireless network . . . . . . . . .
2.1.1 Zero Generation (0G) . . . . . . . . . . .
2.1.2 First Generation (1G) . . . . . . . . . .
2.1.3 Second Generation (2G - 2.75G) . . . . .
2.1.4 Third Generation (3G -3.75G) . . . . . .
2.1.5 Fourth Generation (4G) . . . . . . . . .
2.1.6 Fifth Generation (5G) . . . . . . . . . .
2.2 Motion Prediction Algorithms . . . . . . . . . .
2.2.1 ARMA . . . . . . . . . . . . . . . . . . .
2.2.2 Kalman Filter . . . . . . . . . . . . . . .
2.3 RAT selection for multi-homed mobile terminals
2.3.1 MADM . . . . . . . . . . . . . . . . . .
2.3.2 Policy-based . . . . . . . . . . . . . . . .
i

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.

1
1
2
2
2
2
3

.
.
.
.
.
.
.
.
.
.
.
.
.

4
4
4
5
5
6
7
8
10
10
12
15
15
21

3 System Modelling
3.1 Arrival Rate Modelling . . . . . . . . . . . . . . . . .
3.2 Usage Time Modelling . . . . . . . . . . . . . . . . .
3.3 Network Infrastructure Modelling . . . . . . . . . . .
3.3.1 Physical attributes of the RAT . . . . . . . .
3.3.2 Coverage of Network . . . . . . . . . . . . . .
3.3.3 Capacity and CAC . . . . . . . . . . . . . . .
3.4 Traffic and Utility Function Modelling . . . . . . . .
3.5 Signal Strength Modelling . . . . . . . . . . . . . . .
3.6 Battery Modelling . . . . . . . . . . . . . . . . . . . .
3.7 User Movement Estimation and Prediction Modelling
3.7.1 Straight Line Movement . . . . . . . . . . . .
3.7.2 Turning Movement . . . . . . . . . . . . . . .
3.7.3 Controlled Movement . . . . . . . . . . . . . .
3.7.4 Implementation of Kalman Filter . . . . . . .
3.7.5 Prediction Model Validation . . . . . . . . . .
4 Design
4.1 Design Parameters . . . . . . . . . . . .
4.2 Algorithm Design . . . . . . . . . . . . .
4.2.1 Candidate Criterion . . . . . . .
4.2.2 Bandwidth Allocation Algorithm
4.2.3 MADM Algorithm . . . . . . . .
4.2.4 RAT Connection Algorithm . . .
4.2.5 Overview . . . . . . . . . . . . .
4.2.6 Software Structure . . . . . . . .
5 Results & Discussion
5.1 Validation of Implementation . .
5.1.1 Experiment 1 . . . . . . .
5.1.2 Experiment 2 . . . . . . .
5.2 Benefits of Prediction Mechanism
6 Conclusion

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

23
23
23
24
24
24
26
27
29
30
31
31
32
33
33
33

.
.
.
.
.
.
.
.

40
40
41
41
45
47
49
51
53

.
.
.
.

55
55
55
59
63
64

ii

7 Future Work
7.1 Road Network System . . . . . . . . . . . . . . . . . . . . . . . .
7.2 Dynamic Adjustment of Bandwidth Allocated to Elastic Services .
7.3 Re-scaling Bandwidth Allocated to Elastic and Adaptive Services
7.4 Including More Attributes in Comparison . . . . . . . . . . . . . .
7.5 Including Predictions in Comparison . . . . . . . . . . . . . . . .

.
.
.
.
.

65
65
65
66
66
66

References

69

A Simulation software

70

iii

List of Figures
2.1
2.2
2.3
2.4

Graphical representation of OLS method [1] . . . . . . . . . . . .
Graphical representation of process of Kalman Filter [2] . . . . . .
Detailed Graphical representation of process of Kalman Filter [2] .
Policy based Interface Selection Mechanism [3] . . . . . . . . . . .

.
.
.
.

12
13
15
22

3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
3.14

Coverage of a transceiver . . . . . . . . . . . . . . . . . . . . . .
CAC policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Flow chart of CAC . . . . . . . . . . . . . . . . . . . . . . . . .
Utility functions of different applications [4] . . . . . . . . . . .
Vector Representation of Straight Line Movement . . . . . . . .
Vector Representation of Turning Movement . . . . . . . . . . .
Error when measurement noise is 10 and sampling period is 0.5
Error when measurement noise is 10 and sampling period is 0.1
Error when measurement noise is 100 and sampling period is 0.5
Error when measurement noise is 100 and sampling period is 0.1
Error of prediction and actual position . . . . . . . . . . . . . .
Error of prediction and actual position . . . . . . . . . . . . . .
Error of prediction and actual position . . . . . . . . . . . . . .
Error of prediction and actual position . . . . . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.

25
26
27
29
32
32
34
35
35
36
37
38
38
39

4.1
4.2
4.3
4.4
4.5

Uncertainty Radius . . . . . . . .
Impact of parameters on selection
Connection Algorithm Overview .
System Overview . . . . . . . . .
Class Diagram of the simulation .

.
.
.
.
.

.
.
.
.
.

43
44
50
52
54

5.1

Simulation results of experiment 1 . . . . . . . . . . . . . . . . . . . 56
iv

. . . . . .
mentality
. . . . . .
. . . . . .
. . . . . .

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

5.2
5.3
5.4

Graphical representation of results of experiment 1 . . . . . . . . . 57
Simulation results of experiment 2 . . . . . . . . . . . . . . . . . . . 60
Diagram of simulation results of experiment 2 . . . . . . . . . . . . 61

v

List of Tables
I
II
III

Specifications of different mobile networks [5] . . . . . . . . . . . . . 9
MADM Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Different normalisation methods . . . . . . . . . . . . . . . . . . . . 17

I
II

Different types of traffic [4] . . . . . . . . . . . . . . . . . . . . . . . 28
Relationship between parameters and accuracy of movement model 36

I
II
III
IV
V
VI
VII
VIII
IX

Relevant information of RATs . . . . . . . . . . . . .
Relevant information of user . . . . . . . . . . . . . .
Decision Matrix at position {62,53} . . . . . . . . . .
Cj index of different combinations at {62,53} . . . . .
Decision Matrix at position {82,42} . . . . . . . . . .
Cj index of different combinations at position {82,42}
Relevant information of RATs . . . . . . . . . . . . .
Relevant information of user . . . . . . . . . . . . . .
Cj index at position {0,15} . . . . . . . . . . . . . . .

vi

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

56
56
57
58
59
59
60
60
62